Disclosed is a method of training at least a part of a neural network including a plurality of layers performed by a computing device according to an exemplary embodiment of the present disclosure. The method includes: inputting training data including normal data and abnormal data to an input layer of the neural network; making a feature value output from each of one or more hidden nodes of a hidden layer of the neural network for each training data into a histogram and generating a distribution of the feature value for each of the one or more hidden nodes; calculating an error between each distribution of the feature value and a predetermined probability distribution; and selecting at least one hidden node among the one or more hidden nodes of the hidden layer based on the error.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method of training at least a part of a neural network including a plurality of layers, the method being performed by a computing device, the method comprising: inputting training data including normal data and abnormal data to an input layer of the neural network; making a feature value output from each of one or more hidden nodes of a hidden layer of the neural network for each training data into a histogram and generating a distribution of the feature value for each of the one or more hidden nodes; calculating an error between each distribution of the feature value and a predetermined probability distribution; and updating a weight of at least one hidden node to deactivate the at least one hidden node of the one or more hidden nodes of the hidden layer based on the error.
2. The method of claim 1 , wherein an initial weight of the neural network including the plurality of layers is randomly determined.
3. The method of claim 1 , wherein the selecting of at least one hidden node among the one or more hidden nodes of the hidden layer based on the error includes: selecting a hidden node, in which the error is equal to or smaller than a predetermined value, among the one or more hidden nodes of the hidden layer.
4. The method of claim 1 , wherein the neural network includes at least three layers of the hidden layer.
5. The method of claim 1 , wherein the training data does not include labelling and the training method is an unsupervised training method.
6. The method of claim 1 , wherein the predetermined probability distribution is a Weibull distribution, in which a parameter is randomly determined.
7. The method of claim 1 , further comprising: normalizing the feature value output from the hidden layer.
8. The method of claim 1 , further comprising: inputting the training data to an input layer of each of a plurality of neural networks.
9. The method of claim 8 , further comprising: reconfiguring the neural network by making an ensemble of one or more nodes selected from the plurality of neural networks.
10. A computer program stored in a non-transitory computer readable storage medium, the computer program including a plurality of commands executed by one or more processors of a computing device, the computer program comprising: a command for inputting training data including normal data and abnormal data to an input layer of a neural network; a command for making a feature value output from each of one or more hidden nodes of a hidden layer of the neural network for each training data into a histogram and generating a distribution of the feature value for each of the one or more hidden nodes; a command for calculating an error between each distribution of the feature value and a predetermined probability distribution; and a command for updating a weight of at least one hidden node to deactivate the at least one hidden node of the one or more hidden nodes of the hidden layer based on the error.
11. A computing device for training at least a part of a neural network including a plurality of layers, the computing device comprising: one or more processors; and a memory, which stores commands executable by the one or more processors, wherein the one or more processors input training data including normal data and abnormal data to an input layer of the neural network, make a feature value output from each of one or more hidden nodes of a hidden layer of the neural network for each training data into a histogram and generate a distribution of the feature value for each of the one or more hidden nodes, calculate an error between each distribution of the feature value and a predetermined probability distribution, and update a weight of at least one hidden node to deactivate the at least one hidden node of the one or more hidden nodes of the hidden layer based on the error.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
July 30, 2018
September 8, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.